Summary
When setting up a virtual Exinda, there might arise questions as to what the underlying physical hard disk structure should be in order to get the best performance from the Exinda VM.Overview
When setting up a virtual Exinda, the hypervisor hardware that it sits on can allocate portions of hardware dedicated to the VM. Components such as RAM and virtual CPU (vCPU) are easy to allocate in hypervisor. When creating a virtual hard disk (VHD in Hyperv-V and Citrix XenServer, VMDK in ESXi), it will use whatever datastore it is pointed at, whether those collections of hard drives are stored on network attached storage (NAS) or on a storage area network (SAN), or local.With components such as RAM, vCPU and virtual NICs, there are very few choices how to implement them as any underlying hardware is standard - a CPU that runs at the clock speed required with multiple cores will work fine. However, with hard drives, there are a lot of different physical types - SSD, SAS, iSCSI or Fiber Channel running from a SAN, and all at various RPM. These different connection technologies and speeds can cause slight changes in the level of performance of the drive compared to others.
A virtual hard disk is a file that is located on the physical disk, so it is accessed in a similar way as any other type of file on a disk. Things like read time, write time, seek time, are all important to take into consideration, especially considering that the Exinda is a very I/O heavy system - every packet through the appliance is recorded in some way, whether it is to the database, or information passed through processes. For high bandwidth traffic, this equates to a large number of reads and writes on drives. However, none of the processes require any sort of 'special' features, even at heavy load.
Exinda Hardware appliances come with SAS or SATA 7200RPM drives. As a result, any SAS / SATA 7200RPM drive or greater is known to work with virtual Exindas as underlying hardware. SSDs as well (that do not require any RPM) will work as well, though some SSDs have a limited number of writes that can be performed successfully before fail. This number has increased dramatically over time, and will take years at constant, sustained use.